adaplmmintervals: Prediction intervals.

In schaarschmidt/mixADA: Normalization, mixture models and screening cutpoints for anti-drug-antibody reactions

Description

For internal use in the mixADA GUI: Fits simple random effects models (using package lme4) for a chosen subset of bioogical samples (acc. to previous classification in a mixture model), and computes prediction intervals based on the model fit. Also, percentile intervals for the subsets resulting from previous mixture model fits are returned.

Usage

adaplmmintervals(resadapmixmod, design = c("c2", "h2", "c1", "h1", "y"), level = 0.95, 
alternative = "less", group = c("nonresponder", "responder", "all"))

Arguments

resadapmixmod	an object obtained by fitting a mixture model, by using functions adapmixmod
design	one of three character strings naming an experimental design: "c2" invokes fitting a two-factor random effects model with interaction, "h2" invokes fitting a two-factor hierarchical random effects model, "c1" invokes fitting a two-factor random effects model without interaction, "h1" invokes fitting a one-factor hierarchical random effects model (biological samples and nested in these, repeated measurements), "y" invokes a simple model with only one variance for the biological samples
level	single numeric value, ]0,1[, the levels of prediction limit
alternative	as usual, single character string, "two.sided": two-sided prediction intervals, "less": upper prediction limits only, "greater": lower prediction limits only
group	a single character string, naming for which subset of biological samples to consider for fitting random effects models and prediction limits

Value

a list:

estlimitsd	a data.frame with the estimated prediction limits or percentiles, incl some labelling
TAB	a table with sample sizes of biological samples within each technical unit, as used for random effects models
PIE	a list, with detailed results of the prediction limit computation
DAT	the data.frame used for model fitting (by default, only the subset classified as non-responder in the mixture model classification)

and further elements as input

schaarschmidt/mixADA documentation built on May 29, 2019, 3:25 p.m.